Sync the skill with the shipped CLI (v0.0.27)

lium/SKILL.md

@@ -134,14 +134,9 @@ export PATH="$HOME/.lium/bin:$PATH"  # needed in current shell session
 
 After completing the two-step auth, run `lium ls` to verify. If it returns results, auth is done.
 
-#### `lium ps` Has No `--format json`
+#### Target Pods by Name, Not by Row Number
 
-```bash
-lium ps --format json  # Error: No such option
-lium ps                # correct
-```
-
-Use the pod **name** (e.g. `lunar-lion-4c`) from `lium ps` output for targeting — not a numeric index.
+Use the pod **name** or huid (e.g. `lunar-lion-4c`) from `lium ps` output — not a numeric index. Indices shift as pods come and go, so the number an agent read a minute ago may point at a different pod now.
 
 #### Commands Without -y Flag
 
@@ -261,10 +256,10 @@ If an executor is visible on the lium.io dashboard but `lium up <executor_id>` o
 
 ```bash
 lium ls                        # all available GPUs (shows table with ★ for best price/perf)
-lium ls H100                   # filter by type
-lium ls --sort download        # sort by download speed (fastest first) — preferred default
+lium ls --gpu H100             # filter by type (there is no positional `lium ls H100`)
+lium ls --sort download        # sort by download speed (fastest first) — this is the default
 lium ls --sort upload          # sort by upload speed
-lium ls --sort price_gpu       # sort by price per GPU/hour (default)
+lium ls --sort price_gpu       # sort by price per GPU/hour
 lium ls --format json          # machine-parseable output
 lium templates                 # list Docker templates
 lium templates pytorch         # search templates
@@ -276,7 +271,7 @@ lium templates pytorch         # search templates
 
 ```bash
 lium up --gpu H100 -y          # create pod
-lium ps                        # list active pods (no --format json support)
+lium ps                        # list active pods
 lium ssh my-pod                # SSH into pod
 lium exec my-pod "nvidia-smi"  # run command
 lium exec all "pip install torch"  # batch exec on all pods
@@ -315,9 +310,17 @@ Always use `--format json` when parsing output programmatically:
 
 ```bash
 lium ls --format json | python -c "import json,sys; print(json.load(sys.stdin))"
+lium ps --format json | jq '.[] | {huid, status, ports}'
 ```
 
-Note: `lium ps` does NOT support `--format json`.
+Which command takes what (verified against v0.0.27):
+
+| Command | Flag |
+|---|---|
+| `lium ls`, `lium ps` | `--format table\|json` — no other format, `csv` does not exist |
+| `lium balance`, `lium fund`, `lium topup currencies`, `lium topup create` | `--json` |
+| `lium provider …` | `--json` on the group, so every provider subcommand takes it |
+| everything else (`templates`, `volumes`, `schedules`, `ssh-keys`, `exec`, `logs`, `bk`, …) | no machine-readable output — parse the text or use the SDK |
 
 ## End-to-End Agent Workflow
 
@@ -345,8 +348,8 @@ lium ls --gpu H100 --sort download
 # 5. Create pod (non-interactive!)
 lium up --gpu H100 --name work-pod --ttl 6h -y
 
-# 6. Wait and verify (note: --format json is NOT supported for lium ps)
-lium ps
+# 6. Wait and verify
+lium ps --format json
 
 # 7. Use the pod
 lium scp work-pod ./code.py